A list of Hacker News's undocumented features and behaviors

Original Article ↗ Hacker News Discussion ↗

Hacker News users catalogue a wide range of undocumented features, heuristics, and moderation behaviors that shape how stories and comments surface on the site. They highlight karma thresholds, hidden views and lists (like “classic,” “highlights,” and leaderboards), automatic URL and title rewrites, shadow-style mechanisms such as rate limits and spam/promotional filters, and tools like vouching for [dead] posts. Much of the conversation centers on transparency and fairness in moderation — including shadowbans, vote weighting, and anti-flamewar tuning — and how these choices affect perceived censorship, self-censorship, and the overall quality of debate.

Hidden bans, dead posts, and domain filters

  • Some links from banned domains appear only to the submitter; others show as [dead] in /newest when “showdead” is enabled.
  • Certain users report “shadowbanned” behavior: all submissions or comments auto-[dead], or votes having no effect. Others say this is usually tied to spam/abuse or VPN/spam-filter collisions.
  • There’s an explicit policy against “primarily promotional” accounts; repeatedly posting one domain (especially one’s own) eventually causes the software to filter those submissions.

Voting, flags, and vouching

  • 501 karma is confirmed as the downvote/flag threshold. Flagging power is believed to be weighted by user.
  • Comment vote counts are hidden to avoid “score bickering” and bandwagoning. Users only see counts on their own comments.
  • Only [dead] items (not merely [flagged]) can be “vouched” back into visibility; vouching is available from modest karma levels.
  • Some users allege subtle vote manipulation or “weak shadowbans” via ranking weights; others say HN is aggressive about suppressing orchestrated vote manipulation.

Special lists, pages, and navigation tricks

  • /classic, /best, /active, /highlights, /lists, /leaders, /hidden, /vouched, /flagged, /noobcomments, and various RSS feeds (including /bigrss) are discussed.
  • /classic uses votes from “ancient” accounts (created within HN’s first year).
  • /highlights is a manually curated list of notable comments; users are invited to email suggestions.
  • UI shortcuts: /next and /prev move between sibling comments; /root jumps to the top-level ancestor; /context shows a comment in its thread. Deep replies temporarily hide “reply” links unless opened on their own page.

Rate limiting, flamewar detection, and ranking

  • Users describe a hidden per-account comment cap (“you’re posting too fast”) that can last hours or longer. Some see it as flamewar mitigation; others experience it as opaque punishment.
  • An automatic “flamewar detector” reportedly de-ranks threads with very high comment-to-upvote ratios, sometimes affecting intense but civil discussions.
  • Submission ranking occasionally seems inconsistent with raw points/comments, leading to suspicions of manual score weighting; others claim staff are especially cautious around YC-related content.

Moderation philosophy, transparency, and culture

  • Many praise moderation for keeping discourse relatively civil compared to other platforms.
  • Others criticize “secret” tools: invisible rate limits, ranking penalties, and shadow-style restrictions without explicit notice.
  • A moderator explains that the goal is to protect “the vessel” (curiosity-driven discussion), not enforce a single “truth,” and that penalties are usually tied to guideline violations such as personal attacks or low-signal flamebait.
  • There is discussion of making penalties time-limited and better documented, but limited moderator capacity is noted.

Privacy, permanence, and self-censorship

  • Users worry about long-term deanonymization via linguistic fingerprinting and mass scraping, leading some to consider nuking accounts.
  • Others value persistent archives and argue that easy deletion undermines conversation coherence.